Stable Isotopes and Gut Contents Indicate Differential Resource Use by Coexisting Asp (Leuciscus Aspius) and Pikeperch (Sander Lucioperca)

Received: 21 December 2017 | Revised: 11 April 2018 | Accepted: 13 April 2018 DOI: 10.1111/eff.12414

ORIGINAL ARTICLE

Stable isotopes and gut contents indicate differential resource use by coexisting asp (Leuciscus aspius) and pikeperch (Sander lucioperca)

1Institute of Hydrobiology, Biology Centre of the Czech Academy of Sciences, České Abstract Budějovice, Czech Republic Differential use of habitat and prey resources is an important mechanism that may 2 Norwegian Institute for Nature Research, allow coexistence of sympatric species. Unlike interactions between smaller cyprinid Trondheim, Norway and percid fishes, the resource use by coexisting predatory asp (Leuciscus aspius) and Correspondence pikeperch (Sander lucioperca) is relatively unknown. Here, gut content and stable iso- Mojmír Vašek, Institute of Hydrobiology, Biology Centre of the Czech Academy of tope analyses were used to study ontogenetic dietary shifts and interspecific trophic Sciences, České Budějovice, Czech Republic. niche overlap between asp and pikeperch coexisting in two reservoirs. The hypoth- Email: [email protected] esis that both species show an ontogenetic dietary shift from small invertebrates to Funding information large fish prey, but at the same time use different prey resources to reduce potential Norwegian Financial Mechanism 2009– 2014, Grant/Award Number: 7F14316; competitive interactions, was validated. The isotopic niches of the two predators Grantová agentura České republiky, Grant/ showed no, or only a moderate, degree of overlap (0%–65%). The ontogenetic Award Number: 15-01625S; Horizon 2020 Framework Programme, Grant/Award changes in the degree of interspecific isotopic niche overlap were different in the Number: 677039; Norwegian Institute for two reservoirs, suggesting that trophic segregation can be dynamic and variable Nature Research among systems. Gut contents revealed that small (<100 mm standard length) asp consumed mostly terrestrial invertebrates and emerged aquatic insects, whereas small pikeperch foraged on zooplankton, larval and pupal stages of aquatic insects and fish. Larger individuals (>100 mm) of both species were predominantly piscivorous, with asp consuming more cyprinid prey and pikeperch more percid prey. Coexisting asp and pikeperch populations are able to utilise different prey resources, thereby reducing potential negative competitive interactions.

KEYWORDS dietary ontogeny, foraging strategy, interspecific competition, piscivory, stable isotopes

1 | INTRODUCTION time of activity, the habitat used or the type of prey eaten (Schoener, 1986). A comprehensive review of resource use in fish communities Differential resource use is perceived as an important mechanism by Ross (1986) suggested that niche segregation among coexisting allowing the coexistence of species within ecological communi- species is mainly driven by partitioning of available food resources ties (Chase & Leibold, 2003; Chesson, 2000; Schoener, 1986). This rather than habitat or time segregation. Species coexistence can, view is based on the competitive exclusion principle (Hardin, 1960), however, be influenced also by other mechanisms. For example, sto- which states that species cannot stably coexist unless the utilisa- chastic events (e.g. unpredictable environmental fluctuations) that tion of limiting resources is well differentiated. The segregation of affect demographic attributes of species may result in their coexis- coexisting species can occur along various dimensions such as the tence (Grossman, Ratajczak, Crawford, & Freeman, 1998; Sale, 1978;

Warner & Chesson, 1985). Therefore, one approach to improve our the summer season. In contrast, Římov Reservoir (hereafter Římov; understanding of the ecological mechanisms that determine the co- 48°51′00″N, 14°29′28″E), situated on the Malše River, is a deep existence of species is to examine resource overlap among potential canyon-type lake (Table 1) that is strongly thermally stratified during competitors. Most studies of dietary segregation between coex- the summer season. Both reservoirs have similar water clarity and a isting fish species have focused on adult life stages (e.g. Hodgson, moderately eutrophic trophic status (Table 1). Schindler, & Kitchell, 1997; Schulze, Dörner, Baade, & Hölker, 2012; Due to seasonal water level fluctuations, the littoral zone vegeta- Walker, Kluender, Inebnit, & Adams, 2013; Zaia Alves et al., 2017). tion is poorly developed and submerged macrophytes are practically Ontogenetic variations in the resource use among potential com- missing in both reservoirs. The adult fish community compositions petitors have been examined less frequently (Amundsen et al., are similar in Lipno and Římov, with a dominance of cyprinid species 2003; Davis, Blanchette, Pusey, Jardine, & Pearson, 2012; Werner & (mostly roach Rutilus rutilus, bleak Alburnus alburnus, bream Abramis Gilliam, 1984) although this knowledge is important to fully under- brama and white bream Blicca bjoerkna) accompanied by perch Perca stand the structure and functioning of fish communities. fluviatilis and ruffe Gymnocephalus cernua (Čech et al., 2009; Vašek Asp (Cyprinidae, Leuciscus aspius) and pikeperch (Percidae, Sander et al., 2016). Asp and pikeperch naturally reproduce in both res- lucioperca) are important piscivorous fishes in freshwater communi- ervoirs (Blabolil et al., 2016; Jůza et al., 2013). In Římov, however, ties of western Eurasia where they naturally coexist in large rivers, populations of the two predators are also regularly supported by lakes and reservoirs (Kottelat & Freyhof, 2007; Vašek et al., 2013). stocking with pond-reared fingerlings in autumn (Vašek et al., 2013). Adults reach similar sizes (usually up to 1,000 mm in total length) and prey on small fish (Baruš & Oliva, 1995; Mittelbach & Persson, 1998). 2.2 | Sample collection Juveniles of pikeperch forage on aquatic invertebrates, whereas juveniles of asp may also feed on terrestrial insects fallen on the Fish sampling and treatment were conducted in compliance with water surface (Baruš & Oliva, 1995). The feeding ecology of pike- guidelines from the Experimental Animal Welfare Commission under perch has been explored extensively and thus it is well known that the Ministry of Agriculture of the Czech Republic. Asp, pikeperch this species usually shifts to piscivory in the first summer of its life and their fish prey were sampled from Lipno in August/September (Buijse & Houthuijzen, 1992; van Densen, Ligtvoet, & Roozen, 1996; 2012 and 2013 and from Římov in August 2013 and 2014. Sampling Mittelbach & Persson, 1998). Less is known, however, about the size was carried out with multimesh survey gillnets set overnight in lit- and age at which asp become piscivorous. Moreover, only limited toral, profundal and pelagic zones at four to five different stations attempts have been made to quantitatively characterise the diets of within each reservoir (for details of the gillnet sampling, see Vašek coexisting asp and pikeperch populations (Specziár & Rezsu, 2009). et al., 2016). Additional samples of young-of-the-year (YOY) asp and In general, similar feeding habits (i.e. invertivory followed by pisciv- pikeperch, as well as prey fish, were collected from the littoral and ory) suggest that the two species may interact strongly. Sympatric pelagic zones of both reservoirs using a beach seine net and a trawl populations of asp and pikeperch thus provide a good opportunity respectively (for details of these sampling methods, see Jůza et al., to investigate whether and how the two predators differ in resource 2014). use throughout their lives. Each fish was measured for standard length (mm), and a sam- In this study, gut content (GCA) and stable isotope (SIA) analyses ple of dorsal muscle was dissected and stored at −20°C until pro- were used to explore ontogenetic dietary shifts and niche segrega- cessed for stable isotope analysis. The analysed prey fish included tion between asp and pikeperch co-occurring in two artificial lakes YOY perch, ruffe and roach, and 1-year-old bleak. The digestive (i.e. reservoirs). It was expected that both species undergo an ontogenetic dietary shift from invertebrates to fish prey, but this shift occurs later (i.e. at a larger body size) for asp due to its higher ten- TABLE 1 Basic environmental characteristics of the two reservoirs studied. Mean values for the growing season (May– dency to feed on invertebrates. It was also hypothesised that coex- September) are shown for Secchi depth, total phosphorus and isting asp and pikeperch use different prey resources, but the degree chlorophyll a of trophic segregation diminishes with increasing body size, that is when both species become piscivorous. Characteristic Lipno Římov Year of filling 1960 1978 2 | METHODS Surface altitude (m a.s.l.) 725 471 Surface area (km2) 48.7 2.1 2.1 | Study sites Mean depth (m) 6 16 Maximum depth (m) 22 43 The study was carried out in two reservoirs located in South Bohemia, Hydraulic retention time (days) 244 85 Czech Republic. Lipno Reservoir (hereafter Lipno; 48°37′58″N, Secchi depth (m) 1.9 2.6 14°14′13″E), situated on the upper Vltava River, is a relatively shal- Total phosphorus ( g/L) 25 27 low water body (Table 1). Due to its shallowness and frequent wind μ Chlorophyll a ( g/L) 14 19 action, most of the reservoir area does not thermally stratify during μ VAŠEK et al. | 3 tracts of asp and pikeperch were dissected and preserved in a 10% exoskeleton parts of invertebrate prey) were used for identification formaldehyde solution for later diet analysis. Scales and otoliths of masticated and partially digested prey items. The prey items were were taken and used for age determination following validated subsequently grouped into six categories: (a) crustacean zooplank- methods described by Ruuhijärvi, Salminen, and Nurmio (1996) ton, (b) larval and pupal stages of aquatic insects, (c) emerged aquatic and Krpo-Ćetković, Hegediš, and Lenhardt (2010). To evaluate insects, (d) terrestrial insects, (e) cyprinid fish and (f) percid fish. ontogenetic changes in the short-term diets (based on GCA that represents the recently ingested prey items) and long-term diets 2.4 | Stable isotope analysis (based on SIA that represents the assimilated food sources over several weeks to months) of asp and pikeperch, individuals of both Fish muscle and invertebrate samples were dried at 60°C for 48 hr species were grouped into <100, 100–199, 200–299 and ≥300 mm and ground to a fine powder using either a porcelain mortar or a size classes that corresponded approximately to age categories 0+, mixer mill MM 200 (Retsch GmbH, Haan, Germany). Stable carbon 1+, 2+ and ≥3+ respectively. and nitrogen isotopes and the element (C, N) composition of all sam- At both reservoirs, invertebrate samples for SIA were collected ples were measured using a Europa Scientific elemental analyser from three to four sampling stations and three to four times (June to interfaced with a Europa Scientific 20-20 isotope ratio mass spec- August) during the same summers when the fish were captured. Bulk trometer (Sercon Ltd, Crewe, UK) at the Iso-Analytical Ltd, Crewe, zooplankton was collected from the pelagic zone by taking several UK. Vienna Pee Dee Belemnite and atmospheric N2 were used as vertical hauls through the upper 5 m (Lipno) or 10 m (Římov) of the the international standards for carbon and nitrogen respectively, 13 water column with a 200-μ m mesh plankton net. The live zooplank- while NBS-1577B (powdered bovine liver, δ CV-PDB = −21.60‰, 15 ton was immediately sieved through a 350-μ m mesh and stored fro- δ NAir = 7.65‰) was used as a working standard. NBS-1557B was zen at −20°C. Before preparation for SIA, defrosted samples were calibrated in-house as a secondary reference material and is di- 13 visually inspected using a stereomicroscope. Most of the samples rectly traceable to IAEA-CH-6 (sucrose, δ CV-PDB = −10.43‰) and 15 were dominated by herbivorous crustaceans, mainly cladocerans IAEA-N- 1 (ammonium sulphate, δ NAir = 0.40‰). Isotope ratios in 13 (Daphnia) and calanoid copepods. Bulk samples of macroinverte- each sample were expressed in conventional delta notation (δ C, 15 brates from the littoral zone (<2 m depth) were collected with a kick δ N) as parts per thousand (‰) differences from the international net (mesh size 0.7 mm) and hand-picked from stones. Only nonpreda- standard. The analytical error (standard deviation), estimated from tory organisms (primary consumers) were considered, and they in- replicated runs of the reference material, was less than 0.1‰ for 13 15 cluded mainly trichopteran, ephemeropteran and chironomid larvae, both δ C and δ N. Every fifth sample was run in duplicate, and the and also waterlouse (Asellus aquaticus) and small snails (Lymnaeidae). mean difference ± standard deviation (SD) between replicates was 13 15 All the trichopteran larvae and snails were removed from their cases 0.03 ± 0.04‰ for δ C and 0.06 ± 0.06‰ for δ N. The fish muscle 13 or shells. Bulk samples of terrestrial insects were collected with δ C values were not corrected for lipids due to the generally low a sweep net from the shoreline grasses and shrubs. Adults of the C:N ratios (<3.5) indicating negligible lipid content in the samples aquatic insects (Trichoptera, Ephemeroptera, Odonata and Diptera) (Hoffman, Sierszen, & Cotter, 2015). were excluded when present, and thus, the samples contained adult The relative contributions of different diet sources assimilated insects of purely terrestrial origin, that is Hymenoptera, Hemiptera, by each size class of asp and pikeperch were modelled using the SIAR Coleoptera, Lepidoptera, Brachycera and Orthoptera. All littoral and package in R (Stable Isotope Analysis in R; Parnell, Inger, Bearhop, & 13 15 terrestrial invertebrate samples were stored frozen at −20°C until Jackson, 2010). Inputs to the model were the δ C and δ N values further processed for SIA. of the individual consumers (asp and pikeperch) and the reservoir- 13 15 specific mean ± SD δ C and δ N values of the potential prey resources (Figure A1). In both reservoirs, pelagic zooplankton and 2.3 | Gut content analysis littoral macroinvertebrates did not differ in isotope values (t tests, 13 15 In the laboratory, digestive tracts were opened and the contents were p > 0.05 for both δ C and δ N), and hence, they were grouped examined under a stereomicroscope. Since asp lack a true stomach, as “aquatic invertebrates” for SIAR. Furthermore, the isotope data the contents of the entire gut from the oesophagus to the anus were collected in the subsequent years were pooled because isotope analysed, whereas only stomach contents were analysed from pike- values for major trophic level groups (i.e. predatory fish, prey fish, perch. The total gut or stomach fullness was first visually estimated aquatic invertebrates and terrestrial insects) did not substantially on a percentage scale ranging from empty (0%) to full (100%). The differ between years. Therefore, diet composition for both asp and prey items were identified to the lowest feasible taxonomic level, pikeperch was estimated from three possible diet sources: aquatic and their contribution to the total gut or stomach fullness was then invertebrates, terrestrial insects and fish. Fractionation factors determined by the indirect volumetric method (Hyslop, 1980). In (mean ± SD) between resources and the consumers were assumed 13 15 addition, the number of prey fish individuals discernible to species to be 0.91 ± 1.04‰ for δ C and 3.23 ± 0.41‰ for δ N (Vander level was recorded for each digestive tract. When possible, charac- Zanden & Rasmussen, 2001). Element concentrations (proportions teristic remains (e.g. scales, pharyngeal arches, opercula and other of C and N) directly measured in the prey resources were included bones of fish prey, and head capsules, thoraxes, tail spines and other into the model (Phillips & Koch, 2002). 4 | VAŠEK et al.

Trophic position (TP) of individual asp and pikeperch was esti- and pikeperch. The GCA results indicated significant between- mated from stable isotope data, using the equation described by species differences in the prey compositions (ANOSIM: R = 0.457, Cabana and Rasmussen (1996): p < 0.001), but the diets of asp and pikeperch became more similar with increasing size (Table 2). Small (<100 mm) asp fed on terrestrial TP = ( 15N − 15N )∕3.23 + 2 consumer consumer baseline and emerged aquatic insects, whereas larger asp consumed mostly fish (Figure 1). Correspondingly, small (<100 mm) pikeperch foraged 15 15 15 where δ Nconsumer is the δ N value of asp or pikeperch, δ Nbaseline on zooplankton, larval and pupal stages of aquatic insects and fish, 15 is the δ N value of the baseline organisms (calculated as the av- whereas larger pikeperch were mainly piscivores (Figure 1). Contrary erage value from aquatic invertebrates), 3.23 is the assumed diet- to asp, no terrestrial insects or emerged aquatic insects were found 15 tissue enrichment in δ N per trophic level (Vander Zanden & in pikeperch stomachs. The two species showed contrasting prey Rasmussen, 2001), and the constant 2 refers to the TP of the base- fish compositions, with asp feeding more often on cyprinid prey line organisms. fish and pikeperch feeding mostly on percid fish (Figure 1). When Finally, the isotopic niche widths of each size class of asp and only prey fish discernible to species level were considered, the most pikeperch were calculated as sample size-corrected standard ellipse abundant species found in asp guts were ruffe in Lipno and bleak in areas (SEAC) using the SIBER package in R (Stable Isotope Bayesian Římov, whereas the most abundant species observed in pikeperch Ellipses in R; Jackson, Inger, Parnell, & Bearhop, 2011). SEAC was stomachs were perch and conspecifics in Lipno and ruffe in Římov also used to determine the degree of isotopic niche overlap between (Table 3). The logit-regression models (Figure 2) demonstrated that the two species, using the equation of Stasko, Johnston, and Gunn pikeperch shifted to piscivory at a smaller size than asp, both in (2015): Lipno (parameter estimate ± SE for species effect: 2.4 ± 0.9; Z = 2.6,

% Overlap = [{(area of overlap between SEAC1 and SEAC2) p = 0.009) and in Římov (2.3 ± 0.6; Z = 3.7, p < 0.001). The SIA results confirmed the ontogenetic dietary shifts of asp ×2}∕(SEAC1 + SEAC2)] × 100 and pikeperch to piscivory, as illustrated by the positive nonlinear where SEAC1 and SEAC2 are the ellipse areas calculated from asp and relationship between size and TP (Figure 3, Table 4) and by the pikeperch samples respectively. SIAR estimates showing a shift from invertebrate to fish prey with increasing predator size (Figure 4). In Lipno, small (<100 mm) asp had a significantly lower TP than similar-sized pikeperch, whereas 2.5 | Statistical analysis no between-species differences were observed among larger size Nonparametric one-way analysis of similarities (ANOSIM) was classes (Table 5). An opposite pattern was observed in Římov, where run in PAST ver. 3.19 (Hammer, Harper, & Ryan, 2001) to compare no between-species differences in TP were observed for small asp volumetric proportions of different prey categories in the digestive and pikeperch, whereas larger asp had consistently lower TP as com- tracts of different size classes of asp and pikeperch in the Lipno and pared to similar-sized pikeperch (Table 5). The results from SIAR Římov reservoirs. ANOSIM was based on Bray–Curtis similarity isotopic mixing model suggested that terrestrial insects contributed index, and the one-tailed significance was computed by permutation only little to the long-term diet of all size classes of asp and pike- of group membership with 9,999 replicates. The size at piscivorous perch in Římov but had a relatively high contribution to the long- shift was compared between the species using binomial data of prey term diet of both small (<100 mm) asp (54%) and pikeperch (32%) in fish presence in gut contents (0 = no fish remains in gut, 1 = fish re- Lipno (Figure 4). mains in gut) as the response variable and fish length and species as the predictor variables in logit-regression models. Furthermore, TABLE 2 Sample sizes in gut content analysis (GCA) (n) and results from pairwise one-way analysis of similarities (ANOSIM) the ontogenetic (i.e. size-related) changes in asp and pikeperch TP comparisons of volumetric prey proportions in digestive tracts of were analysed by fitting asymptotic regression models using the asp and pikeperch in the Lipno and Římov reservoirs. Statistically SSasymp function in R (Ritz, Baty, Streibig, & Gerhard, 2015). The significant differences (p < 0.05) are shown in bold differences in TP between asp and pikeperch of each size class in n ANOSIM each reservoir were also compared using t test. Finally, the likelihood Size class test in the SIBER (Jackson et al., 2011) was used to test for between- Reservoir (mm) Asp Pikeperch p species differences in isotopic niche widths of asp and pikeperch size Lipno <100 16 17 <0.001 classes. All statistical analyses except ANOSIM were performed in 10 0–199 4 7 0.024 the R computing programme ver. 3.4.1 (R Core Team, 2017). 200–299 6 11 0.262 ≥300 7 6 0.217 Římov <100 12 20 <0.001 3 | RESULTS 10 0–199 14 10 0.043 200–299 15 11 <0.001 Both GCA and SIA data demonstrated clear ontogenetic dietary shifts and differential use of the prey resources by coexisting asp ≥300 7 5 0.176 VAŠEK et al. | 5

FIGURE 1 Mean volumetric proportion of different prey categories in the digestive tracts of different size classes of asp and pikeperch in the Lipno and Římov reservoirs. Predators that contained unidentified prey fish in their digestive tracts are not shown to better illustrate interspecific differences in piscivorous foraging on percid and cyprinid fish. The number of examined digestive tracts with discernible prey items is indicated above the bars

TABLE 3 List of fish species preyed upon and their total pikeperch overlapped partially (i.e. 49–65%; Table 5). An oppo- numbers found in all digestive tracts of asp and pikeperch collected site pattern was observed in Římov, where the interspecific SEAC from the Lipno and Římov reservoirs overlap was highest (65%) in the <100 mm size class, intermediate Asp Pikeperch (42%) in the 100–199 mm size class and none in the 200–299 and ≥300 mm size classes (Table 5). Prey fish species (family) Lipno Římov Lipno Římov

Perch (Percidae) 3 2 14 8 Pikeperch (Percidae) 1 – 11 2 4 | DISCUSSION Ruffe (Percidae) 7 – 6 13 Bleak (Cyprinidae) 3 9 1 – Both GCA and SIA suggested that there were clear ontogenetic di- Bream (Cyprinidae) – – – 1 etary shifts and interspecific niche segregation between asp and Roach (Cyprinidae) 1 1 – 6 pikeperch in the two reservoirs examined. So far, only limited in- Total number of prey fish 15 12 32 30 formation on resource use has been available for co-occurring asp discernible to species level and pikeperch populations (Specziár & Rezsu, 2009), and the current study is also the first that applied SIA approach (together with The isotopic niche widths generally did not differ between the conventional GCA method) to characterise dietary niches of the two coexisting asp and pikeperch populations (Table 5). In Lipno, there sympatric predators. Consequently, findings of this study provide was no overlap between isotopic niches (SEAC) of small (<100 mm) important insights into the trophic ecology of coexisting asp and asp and pikeperch, whereas the isotopic niches of larger asp and pikeperch populations and their roles in freshwater food webs. 6 | VAŠEK et al.

100–199 mm size class attained mean TP values of ≥3.5, indicating 4.1 | Ontogenetic dietary shifts in asp and pikeperch piscivory. Consequently, both species can be characterised as “spe- Gut content and stable isotope data both showed that asp and pike- cialist piscivores” (sensu Keast, 1985) because they shift to piscivory perch consumed more fish prey with increasing body size. According relatively early in life. to the GCA and SIAR results, fish prey overwhelmingly dominated in Gut content analyse indicated that small-sized (<100 mm) asp the short- and long-term diets of large- and medium-sized predators, consumed mainly terrestrial and emerged aquatic insects in both whereas they contributed only around 50% or less to the diets of reservoirs. SIAR results suggested that terrestrial insects dominated small-sized (<100 mm) asp and pikeperch. Our results demonstrate (54%) the long-term diet of small-sized (<100 mm) asp in Lipno, that asp and pikeperch can begin feeding on fish as early as their whereas small asp in Římov showed a greater reliance upon prey first summer (i.e. as YOY), although shifting to piscivory was com- fish. Hence, the SIAR results indicate that small-sized Římov asp pleted in their second summer of life (i.e. at length >100 mm). Both probably consumed more fish than suggested by GCA which reflects the logit-regression models (based on absence/presence of prey fish only recently ingested prey items (e.g. Paradis, Bertolo, & Magnan, in predators’ digestive tracts) and SIA-based TP estimates consist- 2008). According to GCA, small-sized (<100 mm) pikeperch fed on ently indicated that pikeperch shifted to piscivory at a smaller size zooplankton, aquatic insects and fish, but not on terrestrial and than asp. These results confirmed the expectation that juvenile asp emerged aquatic insects. Although terrestrial insects were absent have a higher tendency to feed on invertebrates and shift to pis- in pikeperch digestive tracts, we always included terrestrial inverte- civory somewhat later (i.e. at a larger size) than pikeperch. However, brates as a third prey source (besides aquatic invertebrates and fish) although piscivory occurred later for asp, the TP estimates suggest to make the SIAR analysis consistent between asp and pikeperch. that both species accomplished shifting to predominantly piscivo- Consequently, the SIAR results confirmed that terrestrial insects rous feeding in their second summer of life since individuals of the generally represented an unimportant prey source for pikeperch of

FIGURE 2 Logistic regressions showing the probability of FIGURE 3 Trophic positions of asp (n = 43 & 77) and pikeperch finding fish remains in gut contents as a function of asp (n = 43 & (n = 62 & 88) in the Lipno and Římov reservoirs as a function of 65) and pikeperch (n = 56 & 69) standard length in the Lipno and standard length. Lines indicate predicted values for the nonlinear Římov reservoirs (asymptotic) regression models (see parameter estimates in Table 4) VAŠEK et al. | 7

TABLE 4 Parameter estimates and Reservoir Species Parameter Estimate SE t p corresponding t- and p-values for the nonlinear (asymptotic) regression models Lipno Asp Asym 3.85 0.08 45.79 <0.001 with trophic position (TP) as a response R0 0.22 0.88 0.25 0.806 variable and standard length (mm) as a lrc −4.29 0.27 −16.00 <0.001 predictor variable, given for asp and pikeperch in Lipno and Římov reservoirs Lipno Pikeperch Asym 3.77 0.05 81.09 <0.001 R0 2.03 0.26 7.93 <0.001 lrc −4.16 0.20 −21.28 <0.001 Římov Asp Asym 3.91 0.07 54.77 <0.001 R0 2.50 0.39 6.35 <0.001 lrc −4.45 0.35 −12.90 <0.001 Římov Pikeperch Asym 4.31 0.08 51.30 <0.001 R0 2.86 0.18 16.07 <0.001 lrc −4.66 0.24 −19.16 <0.001

The models are fitted using SSasymp function in R (R Core Team 2017), producing estimates for the horizontal asymptote on the right side (Asym), the response value (i.e. TP) when length is zero (R0), and the natural logarithm of the rate constant (lrc). See Figure 3 for sample sizes and predicted regression curves.

FIGURE 4 Proportional contribution (mean ± 95% credibility intervals) of different prey types (aquatic invertebrates, terrestrial insects and fish) to the diet of different size classes of asp and pikeperch in the Lipno and Římov reservoirs, estimated using SIAR mixing model. The number of predators analysed for stable isotopes in each size class is indicated across the top of the graphs all size classes. Small-sized (<100 mm) Lipno pikeperch were, how- apparent bias (i.e. overestimation of the terrestrial prey contribution ever, an exception because the SIAR suggested that terrestrial in- to pikeperch diet) can be explained by the fact that stable isotope sects might be a substantial diet source (32%) for these fish. This values of terrestrial and aquatic invertebrates partially overlapped 8 | VAŠEK et al.

TABLE 5 Mean trophic position (TP; standard deviation in parentheses), isotopic niche width measured as standard ellipse area (SEAC) and niche overlap for asp and pikeperch of four size classes collected from the Lipno and Římov reservoirs

TP (SD) SEAC Size class Overlap Reservoir (mm) Asp Pikeperch p* Asp Pikeperch p** (%)

Lipno <100 2.5 (0.3) 3.1 (0.2) <0.001 2.6 3.9 0.88 0.0 10 0–199 3.5 (0.2) 3.6 (0.3) 0.64 3.0 2.9 0.38 48.8 200–299 3.7 (0.3) 3.7 (0.2) 0.94 3.2 1.3 0.03 54.9 ≥300 3.8 (0.2) 3.8 (0.2) 0.64 2.1 1.1 0.08 64.7 Římov <100 3.3 (0.3) 3.5 (0.3) 0.07 2.7 2.7 0.51 65.2 10 0–199 3.7 (0.2) 3.9 (0.1) <0.001 1.0 1.2 0.74 42.3 200–299 3.8 (0.1) 4.2 (0.2) <0.001 1.2 1.1 0.28 0.0 ≥300 3.9 (0.2) 4.2 (0.2) <0.001 1.4 0.9 0.18 0.0

p* and p** refer to statistical significances for TP comparisons and for niche width comparisons respectively between asp and pikeperch in each size class. Statistically significant differences (p < 0.05) are highlighted in bold. For number of predators analysed for stable isotopes in each size class, see Figure 4.

(particularly in terms of 13C) in Lipno, hindering assessment of the δ 4.2 | Interspecific niche segregation between relative contributions of these prey sources to higher trophic levels. asp and pikeperch Nevertheless, the SIAR results appropriately revealed the ontogenetic niche shift from feeding on invertebrates to piscivory in both The isotopic niche widths generally did not differ between the co- species. existing asp and pikeperch populations, indicating a similar extent Previous studies have shown that, under favourable growth con- of trophic specialisation in both species. Moreover, the isotopic ditions (i.e. higher optimum temperature and food availability), pike- niches of the two predators showed no or only a moderate degree perch become piscivorous during their first summer and reach sizes of overlap (i.e. 0%–65%). Interestingly, overlap between the isotopic well above 100 mm (Buijse & Houthuijzen, 1992; van Densen et al., niches of asp and pikeperch increased with increasing predator size 1996). In contrast, under less suitable conditions, YOY pikeperch in Lipno but decreased in Římov. These findings do not support the either remain invertivorous and reach generally small sizes (Ginter, hypothesis that the degree of trophic segregation between asp and Kangur, Kangur, Kangur, & Haldna, 2011; Specziár, 2005; Vinni, pikeperch should diminish with increasing body size (i.e. with a shift Lappalainen, Malinen, & Lehtonen, 2009) or develop a bimodal size to piscivory). Instead, the results suggest that size-related trophic distribution with a minor group becoming piscivorous and a majority segregation between asp and pikeperch may be dynamic and vari- staying invertivorous (van Densen, 1985; Frankiewicz, Dąbrowski, & able among systems, probably reflecting varying availability of prey Zalewski, 1996). Information on ontogenetic dietary shifts in asp is sources. limited. Yet, the data available from Lake Balaton (Specziár & Rezsu, The results also demonstrate that coexisting asp and pikeperch 2009) correspond well with the current study: the <40 mm asp were used rather different prey resources both at small and at large sizes. invertivorous, the 41–120 mm asp had a diet containing both inver- The GCA and SIA data both indicated that the interspecific tro- tebrates and fish, and the 121–500 mm fish were entirely piscivo- phic segregation in the smallest (<100 mm) predator size class was rous. The fact that transition to piscivory in our study systems was likely due to the exclusive utilisation of terrestrial invertebrates not completed during the first summer implies growth-limiting con- and emerged aquatic insects by asp, whereas pikeperch used zoo- ditions for juvenile stages of both species. Persson and Brönmark plankton, larval and pupal stages of aquatic insects, and small fish. (2002) highlighted the importance for YOY predators to be syn- Similarly, Specziár and Rezsu (2009) observed that small (16–40 mm) chronised with fluctuations in resource availability. Hence, we can asp foraged mostly on adult Chironomidae, whereas co-occurring speculate that discontinuous availability of suitable food resources similar-sized pikeperch relied on zooplankton. Moreover, the GCA might reduce growth and delay shifting to piscivory in our study sys- results showed that piscivorous stages of the two predators con- tems. However, growth rates of YOY predators might have also been sumed the same fish species, but in different proportions. The fact restricted by water temperatures. Římov is a deep reservoir situated that asp used relatively more cyprinid prey fish while pikeperch con- in a canyon, and hence, it warms slowly in spring which may delay sumed more percid fish might be another reason for the observed the spawning period and shorten the first-year growth season (cf., segregation of the isotopic niches of the two predators, particularly Jůza et al., 2013; Wysujack, Kasprzak, Laude, & Mehner, 2002). In those of medium- and large-sized classes. Because pikeperch ingest contrast, Lipno is a shallow reservoir, but because of its location at a prey intact, with no mastication, it was usually possible to identify higher altitude, thermal conditions may delay spawning and reduce (at least to family level) most of prey fish. In contrast, ingested prey the growth of juvenile stages similarly to that in Římov. fish in asp were often strongly masticated by pharyngeal teeth and VAŠEK et al. | 9 digested beyond recognition. Hence, we suppose that the relative Financial Mechanism 2009–2014 under contract number MSMT- contribution of cyprinid prey fish in the diet of asp might even be 28477/2014 and No. 677039 (ClimeFish) of the European Union’s higher than suggested by the GCA, because small and soft cyprinid Horizon 2020 research and innovation programme. The study was species such as bleak were probably under-represented due to their also partly supported by internal funds of the Norwegian Institute rapid digestion. for Nature Research. This study provides novel empirical data on piscivorous diets of coexisting asp and pikeperch populations. Previous single-species ORCID studies indicated that the piscivorous diets of both asp and pikeperch are dominated by cyprinid (Krpo-Ćetković et al., 2010; Specziár, Mojmír Vašek http://orcid.org/0000-0001-6386-4015 2011; Wysujack et al., 2002) and by percid prey fish (Frankiewicz, Dąbrowski, Martyniak, & Zalewski, 1999; Keskinen & Marjomäki, 2004; Vostradovský & Váša, 1981). Hence, both predators can be- REFERENCES have rather opportunistically and consume the most abundant fish Amundsen, P. A., Bøhn, T., Popova, O. A., Staldvik, F. J., Reshetnikov, Y. species. However, in sympatry, asp and pikeperch can differentiate S., Kashulin, N. A., & Lukin, A. A. (2003). Ontogenetic niche shifts and resource partitioning in a subarctic piscivore fish guild. Hydrobiologia, prey fish resources as illustrated by this study. In summary, using a 497, 109–119. https://doi.org/10.1023/A:1025465705717 combination of GCA and SIA, our study indicates that coexisting asp Baruš, V., & Oliva, O. (1995). Fauna of the Czech and Slovak Republics. and pikeperch populations can use different prey resources at both Fishes – Osteichthyes, Vol 28/2. Praha, Czech Republic: Academia (in juvenile and adult life stages, thereby reducing the potential nega- Czech with English summary). tive competitive interactions (Vanni, Duncan, González, & Horgan, Benndorf, J. (1990). Conditions for effective biomanipulation: Conclusions derived from who-lake experiments in Europe. Hydrobiologia, 2009). 200(201), 187–203. https://doi.org/10.1007/BF02530339 Blabolil, P., Ricard, D., Peterka, J., Říha, M., Jůza, T., Vašek, M., … Kubečka, J. (2016). Predicting asp and pikeperch recruitment in a riverine res- 4.3 | Conclusion and recommendation for ervoir. Fisheries Research, 173, 45–52. https://doi.org/10.1016/j. future studies fishres.2015.08.003 Brabrand, Å., & Faafeng, B. (1993). Habitat shift in roach (Rutilus rutilus) The present study demonstrates that coexisting asp and pikeperch induced by pikeperch (Stizostedion licioperca) introduction: Predation forage at the top of the food webs and thereby play similar func- risk versus pelagic behaviour. Oecologia, 95, 38–46. https://doi. tional roles in lake ecosystems. Notably, the trophic niches of the org/10.1007/BF00649504 Buijse, A. D., & Houthuijzen, R. P. (1992). Piscivory, growth, and size- two predators were relatively well separated, both at juvenile and at selective mortality of age 0 pikeperch (Stizostedion lucioperca). older life stages. The observed niche segregation may help to reduce Canadian Journal of Fisheries and Aquatic Sciences, 49, 894–902. potential interspecific resource competition between coexisting asp https://doi.org/10.1139/f92-100 and pikeperch populations. Cabana, G., & Rasmussen, J. B. (1996). Comparison of aquatic food chains using nitrogen isotopes. Proceedings of the National Academy Asp and pikeperch can induce top-down cascading impacts on of Sciences USA, 93, 10844–10847. https://doi.org/10.1073/ lower trophic levels (Benndorf, 1990; Brabrand & Faafeng, 1993; pnas.93.20.10844 Donabaum, Schagerl, & Dokulil, 1999). They are also popular game Čech, M., Kubečka, J., Prchalová, M., Frouzová, J., Jankovský, M., & fishes for anglers and therefore often stocked into various systems Muška, M. (2009). Fish community survey in the Lipno Reservoir in 2008. České Budějovice, Czech Republic: Technical Report, Biology (e.g. Ruuhijärvi et al., 1996; Vašek et al., 2013; Wysujack et al., 2002). Centre of the Czech Academy of Sciences, Institute of Hydrobiology. Our results are relevant to fisheries management, because they indi- Chase, J. M., & Leibold, M. A. (2003). Ecological niches: Linking classical cate that different use of the prey resources may potentially mitigate and contemporary approaches. Chicago, IL: University of Chicago interspecific competition between co-occurring asp and pikeperch Press. https://doi.org/10.7208/chicago/9780226101811.001.0001 populations. In future studies, comparison of trophic niches of the Chesson, P. (2000). Mechanisms of maintenance of species diversity. Annual Review of Ecology and Systematics, 31, 343–366. https://doi. two species under conditions of allopatry and sympatry could help org/10.1146/annurev.ecolsys.31.1.343 to determine whether relatively low overlap in resource use is the Davis, A. M., Blanchette, M. L., Pusey, B. J., Jardine, T. D., & Pearson, consequence of interspecific competition or different foraging strat- R. G. (2012). Gut content and stable isotope analyses provide com- egies that evolved in the past. plementary understanding of ontogenetic dietary shifts and trophic relationships among fishes in a tropical river. Freshwater Biology, 57, 2156–2172. https://doi.org/10.1111/j.1365-2427.2012.02858.x van Densen, W. L. T. (1985). Piscivory and the development of bi- ACKNOWLEDGEMENTS modality in the size distribution of 0 + pikeperch (Stizostedion lu- We are grateful to FishEcU members (www.fishecu.cz) for their as- cioperca L.). Journal of Applied Ichthyology, 1, 119–131. https://doi. org/10.1111/j.1439-0426.1985.tb00421.x sistance during field sampling, Tomáš Mrkvička for statistical advice van Densen, W. L. T., Ligtvoet, W., & Roozen, R. W. M. (1996). Intra- and Mary J. Morris for editing the English. We also thank David L. cohort variation in the individual size of juvenile pikeperch, Morgan and two anonymous reviewers for their helpful comments. Stizostedion lucioperca, and perch, Perca fluviatilis, in relation to The study was supported by projects No. 15-01625S of the Czech the size spectrum of their food items. Annales Zoologici Fennici, 33, 495–506. Science Foundation, No. 7F14316 (MacFish) of the Norwegian 10 | VAŠEK et al.

Donabaum, K., Schagerl, M., & Dokulil, M. T. (1999). Integrated manage- Paradis, Y., Bertolo, A., & Magnan, P. (2008). What do the empty stom- ment to restore macrophyte domination. Hydrobiologia, 395(396), achs of northern pike (Esox lucius) reveal? Insights from carbon 13 15 87–97. https://doi.org/10.1023/A:1017015200790 (δ C) and nitrogen (δ N) stable isotopes. Environmental Biology Frankiewicz, P., Dąbrowski, K., Martyniak, A., & Zalewski, M. (1999). of Fish, 83, 441–448. https://doi.org/10.1007/s10641-008- Cannibalism as a regulatory force of pikeperch, Stizostedion lu- 9366-2 cioperca (L.), population dynamics in the lowland Sulejow reser- Parnell, A. C., Inger, R., Bearhop, S., & Jackson, A. L. (2010). Source par- voir (Central Poland). Hydrobiologia, 408(409), 47–55. https://doi. titioning using stable isotopes: Coping with too much variation. PLoS org/10.1023/A:1017001803791 One, 5, e9672. https://doi.org/10.1371/journal.pone.0009672 Frankiewicz, P., Dąbrowski, K., & Zalewski, M. (1996). Mechanism of Persson, A., & Brönmark, C. (2002). Foraging capacity and resource establishing bimodality in a size distribution of age-0 pikeperch, synchronization in an ontogenetic diet switcher, pikeperch Stizostedion lucioperca (L.) in the Sulejów Reservoir, Central Poland. (Stizostedion lucioperca). Ecology, 83, 3014–3022. https://doi. Annales Zoologici Fennici, 33, 321–327. org/10.1890/0012-9658(2002)083[3014:FCARSI]2.0.CO;2 Ginter, K., Kangur, K., Kangur, A., Kangur, P., & Haldna, M. (2011). Diet Phillips, D. L., & Koch, P. L. (2002). Incorporating concentration depen- patterns and ontogenetic diet shift of pikeperch, Sander lucioperca dence in stable isotope mixing models. Oecologia, 130, 114–125. (L.) fry in lakes Peipsi and Võrtsjärv (Estonia). Hydrobiologia, 660, https://doi.org/10.1007/s004420100786 79–91. https://doi.org/10.1007/s10750-010-0393-6 R Core Team (2017). R: A language and environment for statistical comput- Grossman, G. D., Ratajczak, R. E., Crawford, M., & Freeman, M. C. (1998). ing. Vienna, Austria: R Foundation for Statistical Computing. Assemblage organization in stream fishes: Effects of environmental vari- Ritz, C., Baty, F., Streibig, J. C., & Gerhard, D. (2015). Dose-response ation and interspecific interactions. Ecological Monographs, 68, 395–420. analysis using R. PLoS One, 10, e0146021. https://doi.org/10.1371/ https://doi.org/10.1890/0012-9615(1998)068[0395:AOISFE]2.0.CO;2 journal.pone.0146021 Hammer, Ø., Harper, D. A. T., & Ryan, P. D. (2001). PAST: Paleontological Ross, S. T. (1986). Resource partitioning in fish assemblages: A review of statistics software package for education and data analy- field studies. Copeia, 2, 352–388. https://doi.org/10.2307/1444996 sis. Palaeontologia Electronica, 4, 9. https://doi.org/10.1007/ Ruuhijärvi, J., Salminen, M., & Nurmio, T. (1996). Releases of pikeperch s10750-017-3159-6 (Stizostedion lucioperca (L.)) fingerlings in lakes with no established Hardin, G. (1960). The competitive exclusion principle. Science, 131, pikeperch stock. Annales Zoologici Fennici, 33, 553–567. 1292–1297. https://doi.org/10.1126/science.131.3409.1292 Sale, P. F. (1978). Coexistence of coral reef fishes – a lottery for liv- Hodgson, J. R., Schindler, D. E., & Kitchell, J. F. (1997). Diet overlap in a ing space. Environmental Biology of Fishes, 3, 85–102. https://doi. piscivore community. Ecology of Freshwater Fish, 6, 144–149. https:// org/10.1007/BF00006310 doi.org/10.1111/j.1600-0633.1997.tb00156.x Schoener, T. W. (1986). Resource partitioning. In J. Kikkawa & D. J. Hoffman, J. C., Sierszen, M. E., & Cotter, A. M. (2015). Fish tissue lip- Anderson (Eds.), Community ecology: Pattern and process (pp. 91–126). id-C: N relationships for correcting δ13C values and estimating Melbourne, Vic.: Blackwell Scientific Publications. lipid content in aquatic food-web studies. Rapid Communications Schulze, T., Dörner, H., Baade, U., & Hölker, F. (2012). Dietary niche par- in Mass Spectrometry, 29, 2069–2077. https://doi.org/10.1002/ titioning in a piscivorous fish guild in response to stocking of an ad- rcm.7367 ditional competitor – the role of diet specialisation. Limnologica, 42, Hyslop, E. J. (1980). Stomach contents analysis – a review of methods 56–64. https://doi.org/10.1016/j.limno.2011.08.001 and their applications. Journal of Fish Biology, 17, 411–429. https:// Specziár, A. (2005). First year ontogenetic diet patterns in two co- doi.org/10.1111/j.1095-8649.1980.tb02775.x existing Sander species, S. lucioperca and S. volgensis in Lake Jackson, A. L., Inger, R., Parnell, A. C., & Bearhop, S. (2011). Comparing Balaton. Hydrobiologia, 549, 115–130. https://doi.org/10.1007/ isotopic niche widths among and within communities: SIBER – s10750-005-5766-x Stable Isotope Bayesian Ellipses in R. Journal of Animal Ecology, 80, Specziár, A. (2011). Size-dependent prey selection in piscivorous pike- 595–602. https://doi.org/10.1111/j.1365-2656.2011.01806.x perch Sander lucioperca and Volga pikeperch Sander volgensis shaped Jůza, T., Mrkvička, T., Blabolil, P., Čech, M., Peterka, J., Vašek, M., & by bimodal prey size distribution. Journal of Fish Biology, 79, 1895– Kubečka, J. (2013). Occurrence of age-0 dwarf pikeperch in late 1917. https://doi.org/10.1111/j.1095-8649.2011.03127.x summer – an overlooked phenomenon in reservoirs. Journal of Fish Specziár, A., & Rezsu, E. T. (2009). Feeding guilds and food re- Biology, 83, 1444–1452. https://doi.org/10.1111/jfb.12229 source partitioning in a lake fish assemblage: An ontogenetic Jůza, T., Vašek, M., Kratochvíl, M., Blabolil, P., Čech, M., Draštík, V., … approach. Journal of Fish Biology, 75, 247–267. https://doi. Kubečka, J. (2014). Chaos and stability of age-0 fish assemblages org/10.1111/j.1095-8649.2009.02283.x in a temperate deep reservoir: Unpredictable success and stable Stasko, A. D., Johnston, T. A., & Gunn, J. M. (2015). Effect of water clar- habitat use. Hydrobiologia, 724, 217–234. https://doi.org/10.1007/ ity and other environmental factors on trophic niches of two sym- s10750-013-1735-y patric piscivores. Freshwater Biology, 60, 1459–1474. https://doi. Keast, A. (1985). The piscivore feeding guild of the fishes in small fresh- org/10.1111/fwb.12581 15 water ecosystems. Environmental Biology of Fishes, 12, 119–129. Vander Zanden, M. J., & Rasmussen, J. B. (2001). Variation in δ N 13 https://doi.org/10.1007/BF00002764 and δ C trophic fractionation: Implications for aquatic food web Keskinen, T., & Marjomäki, T. J. (2004). Diet and prey size spectrum of studies. Limnology and Oceanography, 46, 2061–2066. https://doi. pikeperch in lakes in central Finland. Journal of Fish Biology, 65, 1147– org/10.4319/lo.2001.46.8.2061 1153. https://doi.org/10.1111/j.0022-1112.2004.00500.x Vanni, M. J., Duncan, J. M., González, M. J., & Horgan, M. J. (2009). Kottelat, M., & Freyhof, J. (2007). Handbook of European freshwater fishes. Competition among aquatic organisms. In G. E. Likens (Ed.), Cornol, Switzerland: Publications Kottelat. Encyclopedia of inland waters (pp. 395–404). Oxford, UK: Elsevier. Krpo-Ćetković, J., Hegediš, A., & Lenhardt, M. (2010). Diet and growth of https://doi.org/10.1016/B978-012370626-3.00201-5 asp, Aspius aspius (Linnaeus, 1758), in the Danube River near the con- Vašek, M., Prchalová, M., Peterka, J., Ketelaars, H. A. M., Wagenvoort, fluence with the Sava River (Serbia). Journal of Applied Ichthyology, 26, A. J., Čech, M., … Kubečka, J. (2013). The utility of predatory fish 513–521. https://doi.org/10.1111/j.1439-0426.2010.01456.x in biomanipulation of deep reservoirs. Ecological Engineering, 52, Mittelbach, G. G., & Persson, L. (1998). The ontogeny of piscivory and 104–111. https://doi.org/10.1016/j.ecoleng.2012.12.100 its ecological consequences. Canadian Journal of Fisheries and Aquatic Vašek, M., Prchalová, M., Říha, M., Blabolil, P., Čech, M., Draštík, Sciences, 55, 1454–1465. https://doi.org/10.1139/f98-041 V., … Kubečka, J. (2016). Fish community response to the VAŠEK et al. | 11

longitudinal environmental gradient in Czech deep-valley reser- and Systematics, 15, 393–425. https://doi.org/10.1146/annurev. voirs: Implications for ecological monitoring and management. es.15.110184.002141 Ecological Indicators, 63, 219–230. https://doi.org/10.1016/j. Wysujack, K., Kasprzak, P., Laude, U., & Mehner, T. (2002). Management ecolind.2015.11.061 of a pikeperch stock in a long-term biomanipulated stratified lake: Vinni, M., Lappalainen, J., Malinen, T., & Lehtonen, H. (2009). Efficient predation vs. low recruitment. Hydrobiologia, 479, 169–180. Stunted growth of pikeperch Sander lucioperca in Lake Sahajärvi, https://doi.org/10.1023/A:1021042308649 Finland. Journal of Fish Biology, 74, 967–972. https://doi. Zaia Alves, G. H., Figueiredo, B. R. S., Manetta, G. I., Sacramento, P. A., org/10.1111/j.1095-8649.2009.02181.x Tófoli, R. M., & Benedito, E. (2017). Trophic segregation underlies Vostradovský, J., & Váša, J. (1981). Asp (Aspius aspius L.) as a new fish the coexistence of two piranha species after the removal of a geo- species for rearing. Bulletin VÚRH Vodňany, 17, 10–13. (in Czech with graphic barrier. Hydrobiologia, 797, 57–68. https://doi.org/10.1007/ English summary). s10750-017-3159-6 Walker, R. H., Kluender, E. R., Inebnit, T. E., & Adams, S. R. (2013). Differences in diet and feeding ecology of similar-sized spotted (Lepisosteus oculatus) and shortnose (Lepisosteus platostomus) gars during flooding of a south-eastern US river. Ecology of Freshwater How to cite this article: Vašek M, Eloranta AP, Vejříková I, Fish, 22, 617–625. https://doi.org/10.1111/eff.12066 et al. Stable isotopes and gut contents indicate differential Warner, R. R., & Chesson, P. L. (1985). Coexistence mediated by re- resource use by coexisting asp (Leuciscus aspius) and cruitment fluctuations: A field guide to the storage effect. American Naturalist, 125, 769–787. https://doi.org/10.1086/284379 pikeperch (Sander lucioperca). Ecol Freshw Fish. 2018;00:1–12. Werner, E. E., & Gilliam, J. F. (1984). The ontogenetic niche and species https://doi.org/10.1111/eff.12414 interactions in size-structured populations. Annual Review of Ecology 12 | VAŠEK et al.

APPENDIX fish. All diet sources were corrected for trophic fractionation 13 15 13 15 Figure A1: Biplots of δ C and δ N values for individual asp (cir- using values (δ C = 0.91, δ N = 3.23) from Vander Zanden and cles; n = 43 & 77) and pikeperch (triangles; n = 62 & 88) and their Rasmussen (2001). Because pelagic zooplankton and littoral mac- principal diet sources in the Lipno and Římov reservoirs. Filled roinvertebrates did not differ in their isotope values, they were squares represent mean ± standard deviation for pelagic zoo- merged as “aquatic invertebrates” for SIAR estimates (see plankton, littoral macroinvertebrates, terrestrial insects and prey Figure 4).